On Tue, Feb 26, 2002 at 03:04:06PM -0600, David Champion wrote: > On 2002.02.26, in <[EMAIL PROTECTED]>, > "Jay R. Ashworth" <[EMAIL PROTECTED]> wrote: > > ... and a generic web search engine, running locally on that machine, > > with access to the messages, will serve that purpose fine, and can > > still be hidden behind whatever authentication you like. > > Maybe I'm simply unaware of the generic web search engines that will > run on my web server and subscribe to whatever authentication mechanism > I like, including checking Mailman mailing list passwords for the > particular mailing list a user is trying to search.
I believe you've misunderstood me, here. If you run ht://dig, or something akin to it, *on a machine local to the mailing-list-archive machine, globally authenticated and authorized to retrieve messages from the archiver web server*, and then place *access to that search engine* behind whatever authentication you require for access to the messages themselves*, you've solved your problem, no? > My point is that if the specific authentication I'm looking must be > devised, there's no functional difference to me between placing it in > the archiver and placing it in a separate search product that knows how > my MLM and archiver work, and can talk to them. But the former seems > like a better software design to me, since the archiver is already part > of my MLM. And I disagree, precisely on point: if you have a clean interface to which the search engine can attach (and you do: HTTP), and there's no impedance mismatch (there isn't; the search engine needs to be returning *website* addresses, anyway), then why bundle it inside -- *especially* when there are packaged solutions... and there are. > I'm in search of the ability to index list archives in a way that does > not violate privacy and that my users will understand. It matters more > to me that I can achieve that (without developing the glue myself each > time Mailman changes its authentication interface) than whether my > solution is Righteous. Mailman isn't the issue, so far as I can see; the *archiver* is. Using a packaged websearch engine relieves you of the tie to Piper (which sucks) or Zest (which, I gather, sucks less), or whatever... I can't imagine that hooking the Mailman auth code could be *that* difficult... Cheers, -- jra -- Jay R. Ashworth [EMAIL PROTECTED] Member of the Technical Staff Baylink RFC 2100 The Suncoast Freenet The Things I Think Tampa Bay, Florida http://baylink.pitas.com +1 727 647 1274 "If you don't have a dream; how're you gonna have a dream come true?" -- Captain Sensible, The Damned (from South Pacific's "Happy Talk") _______________________________________________ Mailman-Developers mailing list [EMAIL PROTECTED] http://mail.python.org/mailman/listinfo/mailman-developers
